VM66323: LOSS OF CONTROL IN MEM INIT WITH >64 CPUS

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• During memory initialization, either at system IPL or
  as a result of a SET STORE command, memory is acquired for
  the Available Zone control block.  This control block
  describes a subset of memory and must be aligned on a
  page boundary. Using HCPGETST for a TYPE=ALIGNED
  control block means that a Loss of Control can occur
  in a path that cannot tolerate it, leading to abends.
  
  So far we have seen this loss of control occur on
  systems with greater than 64 logical CPUs defined.
  The abend types seen could vary, as the loss of control
  can affect a variety of code.  Most commonly, PRG004
  or PRG001 abends were seen.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users on any hardware running with >64       *
  *                 logical processors are exposed to this       *
  *                 issue.                                       *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  ****************************************************************
  * RECOMMENDATION: APPLY PTF                                    *
  ****************************************************************
  During memory initialization, either at system IPL or as a
  result of a SET STORE command, memory is acquired for the
  Available Zone control block. As this memory is being
  acquired a Loss Of Control can occur in a path that
  cannot tolerate it. This leads to unpredictable system
  behavior, including PRG001 and PRG004 abends.
  Initializing large amounts of memory with a large number
  of logical processors, especially while also starting
  additional workload, makes this problem more likely.
  IBM recommends adding storage and, only once that is
  complete, starting additional workload.
  

Problem conclusion

• The code was changed to acquire storage without a loss of
  control, which also involved adding logic to delay if acquiring
  storage failed.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

  UM35542

Modules/Macros

• HCPISM   HCPPCQ
  

Fix information

• Fixed component name

  VM CP

• Fixed component ID

  568411202

Applicable component levels

• R710 PSY UM35542

     UP19/09/30 P 2001

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.